Organization repository on GitHub
Deep research tool for local knowledge base.
- Forks
- 46
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Tools & CLIs
DeepDoc
A local deep-research tool that turns your own PDF, DOCX, PPTX, image, TXT, or Markdown files into a structured report, using TypeSafe's Jev to rerank Qdrant chunks and verify evidence coverage while an LLM plans and writes.
- Stars
- 313
- Language
- Python
- License
- MIT
- Reading time
- 2 min
- Published by
- Community
- Author
- Oqura-ai
- Last verified
- Oct 2026
Used inJev for Software & DevTools
- Languages
- Python
- Capabilities
- Open sourceRuns locally
Highlights
- Uses TypeSafe's Jev to check and rerank the chunks returned by Qdrant, without replacing the generative LLM.
- Jev also checks whether collected evidence covers every required subsection; Python uses Jev's probabilities for filtering and retries.
- Extracts text with page and source metadata from PDF, DOCX, PPTX, image, TXT, or Markdown and creates searchable chunks in Qdrant.
- A report planner proposes the outline for approval before research; one research agent then runs per section.
- Section agents use Jev to judge relevance, usable evidence, contradiction, and prompt-injection-like content.
Watch out
README reviewed 2026-10-17 and truncated mid-list in How It Works, so installation, model configuration, and dependencies are not in the digest. It names no Jev model or price, reports no quality measurements, and the topics (ollama, openai, langchain, langgraph) hint at stack choices the excerpts do not document.
Questions about this entry
Short answers drawn from the linked project or page.
A local deep-research tool from Oqura that turns your own files into a structured Markdown report: it extracts and indexes content, plans the report, researches each section against your collection, and combines everything into one document.
Jev checks and reranks the chunks Qdrant returns and verifies that the gathered evidence covers every required subsection. The LLM still handles planning, query generation, synthesis, and writing, while Python uses Jev's probabilities to make filtering and retry decisions.
A directory of PDF, DOCX, PPTX, image, TXT, or Markdown resources. DeepDoc extracts the text, keeps page and source metadata, and creates searchable chunks in Qdrant.
Per the README, each section agent generates focused queries, retrieves candidate chunks, then uses Jev to judge relevance, usable evidence, contradiction, and prompt-injection-like content before filtering and reranking.
Yes. A report planner creates the initial structure from your topic and outline, and you can approve it or request changes before any research agents start.
The tool is positioned as local, and its topics include ollama; the README excerpt does not show configuration or keys, and the digest truncates before dependencies, so check the repo docs before planning a deployment.
Oqura-ai publishes it under MIT; digest metadata shows 312 stars, 46 forks, Python, created 2025-08-15, and a last push on 2026-09-26. A Discord community is linked from the README.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: A RAG loop where Jev decides what is real
X: Measuring what Jev does in a RAG pipeline
X: Jev-powered reranking for RAG. Likitd_’s jev-ranker uses Jev to rerank results in a RAG pipeline.
X: Coverage is not product value
More Jev resources like this
A local experiment bench that uploads PDFs, extracts and chunks their text, then generates Full, Jev and RAG comparison records for consecutive questions, storing workspaces, chunk rules and answer history in a SQLite file at .runtime/jev_context.sqlite3.
Jev Tools & CLIs#community#github#python
Community
wdwd200/jevPython
wdwd200/jevPythonA local workbench that pairs OpenCV vision with Laya or Jev decision models: images are segmented and measured, hard rules settle the numbers, and the model gives the verdict — with one-click installers and an "open-source JEV" training path served at /v1/systemone.
Jev Tools & CLIs#community#github#python
Community
uu889/laya-opencvPython
uu889/laya-opencvPythonFills job applications page by page in a browser from a JSON profile, with plain rules answering common fields, a local decision model choosing among your own facts, and submit gated until you opt in.
Jev Tools & CLIs#community#github#job-applications
Community
Noface07/SystemOne-ApplyPython
Noface07/SystemOne-ApplyPythonA technical knowledge assistant that runs adaptive RAG over 11 technical books (2,810 chunks) with hybrid dense and BM25 search, reciprocal rank fusion and a JEV System-1 decision engine over a local vector database.
Jev Tools & CLIs#community#github#rag
Community
rudra-p11/techbook-ragHTML
rudra-p11/techbook-ragHTMLA self-hosted server that removes ads from podcasts before you hit play: it transcribes locally with faster-whisper or Parakeet, detects and cuts ad segments with your choice of LLM or System One, and learns sponsors across episodes.
Jev Tools & CLIs#community#github#python
Community
ttlequals0/minuspodPython
ttlequals0/minuspodPythonNatural-language lookup over the US App Store top charts: describe an ambiguous feature, a look, or a name and get matching apps with how likely each one is, scored by a single Jev request over a local shortlist.
Jev Tools & CLIs#community#github#appstore
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.